AI Music Remastering

A professional desktop app for mastering AI-generated music to streaming-ready quality.

Features

• Batch Processing - Queue multiple files, apply the same settings, and export all at once
• Metadata Editor - Add title, artist, album, genre, year, track number, and comments per file
• Loudness Normalization - Adjustable target LUFS (-20 to -6 LUFS)
• True Peak Limiting - Prevents clipping with adjustable ceiling
• Input Gain Control - Adjust input level before processing (-12 to +12 dB)
• Stereo Width - Control stereo image (0% mono to 200% extra wide)
• 5-Band EQ - Fine-tune with visual faders and presets (Flat, Vocal Boost, Bass Boost, Bright, Warm, AI Fix)
• Quick Fix Tools - Glue compression, clean low end
• Polish Effects - Cut mud, add air, tame harshness
• Real-time Preview - Hear all changes live before exporting, preview any queued file
• Clipping Detection - Visual CLIP indicators on meters
• High-Quality Export - WAV output at 44.1/48kHz, 16/24-bit with embedded metadata

Download

Get the latest release for your platform:

• Windows - .exe installer
• macOS - .dmg disk image
• Linux - .AppImage

Usage

1. Drag & drop an audio file (MP3, WAV, FLAC, AAC, M4A)
2. Preview with the built-in player
3. Adjust EQ, loudness, and mastering settings
4. Toggle FX bypass to compare before/after
5. Click "Export WAV" for a single file

Batch Processing

1. Click "+ Add Files" or drag multiple files into the batch queue
2. Preview any queued file by clicking the ▶ button to load it into the player
3. Switch to the "Metadata" tab to add tags (title, artist, album, etc.) per file
4. Use "Apply to All" to copy metadata across the entire queue
5. Click "Export All" and choose an output folder

Building from Source

# Install dependencies
npm install

# Build the app
npm run build

# Run in development
npm run electron:dev

# Build for your platform
npm run electron:build:win    # Windows
npm run electron:build:mac    # macOSnpm run electron:build:linux  # Linux

Tech Stack

• Electron + Vite
• Pure JavaScript audio processing (no FFmpeg)
• Web Audio API for real-time preview
• ITU-R BS.1770-4 compliant LUFS measurement
• Native JavaScript WAV encoder

Changelog

v2.0.2

New Features

• Light/dark mode toggle — click the sun/moon button in the top bar to switch themes
• Theme preference is saved to localStorage and persists across restarts

Bug Fixes

• Fixed spectrogram rendering issue caused by devicePixelRatio canvas scaling mismatch

v2.0.1

Bug Fixes

• Fixed metadata tab file list being too narrow to read filenames with many songs queued
• Widened file list panel, increased font size, and expanded scroll area for better readability

v2.0.0

New Features

• Batch processing queue — add multiple files, apply the same mastering settings, and export all at once
• Per-file metadata editor with tabbed UI (title, artist, album, genre, year, track #, comment)
• "Apply to All" button to copy metadata across the entire queue
• WAV metadata embedding via LIST/INFO chunks (INAM, IART, IPRD, IGNR, ICRD, ITRK, ICMT)
• Queue preview — click ▶ on any queued file to load it into the player and audition before exporting
• Files loaded into the player are automatically added to the batch queue
• Single-file "Export WAV" now includes metadata if the file has tags set in the queue
• Multi-file and directory selection dialogs for batch workflows

Improvements

• UI yields to the event loop between batch processing steps to prevent freezing
• Batch progress shows per-file status (pending, processing, done, error) and current filename
• Currently loaded file is highlighted in the queue with a "Loaded" indicator

v1.2.2

Bug Fixes

• Fixed incomplete spectrogram rendering (was truncated mid-function)
• Fixed stereo width being applied twice during export (preview and export now match)
• Fixed seek race condition — isSeeking flag now properly guards playback restart
• Fixed stopAudio ghost callbacks from onended firing after stop

Architecture

• Extracted shared createProcessingNodes() factory for both preview and export chains (DRY)
• Shared configureFilterNodes() accepts settings object — no more duplicated filter setup
• Reduced analyser FFT size to 512 for level meters (faster response)
• Spectrogram throttled to ~30fps with offscreen canvas created once outside draw loop

New Features

• Undo/redo system (Ctrl+Z / Ctrl+Shift+Z) with 50-level history
• Settings persistence via localStorage — all settings and EQ presets survive restarts
• Keyboard shortcuts: Space (play/pause), Escape (stop), B (bypass), ←→ (seek ±5s), Ctrl+E (export)
• Shortcuts hint bar in the UI
• Status messages now appear as floating toasts that auto-dismiss

Audio Quality

• TPDF dithering for 16-bit WAV exports (reduces quantization artifacts)
• Proper 44.1kHz K-weighting filter coefficients for LUFS measurement

Accessibility

• ARIA labels on all interactive controls
• ARIA roles on meters, regions, status areas, and progress bars
• focus-visible outlines for keyboard navigation
• Decorative elements marked aria-hidden

v1.2.0

• Initial release

License

ISC